Just came across Scribble Diffusion!
You can now generate an image based on a quick sketch and a text prompt.
They're using ControlNet, a NN that lets you control pretrained diffusion models and add input conditions. Code: github.com/lllyasviel/Con…scribblediffusion.com
Here are some resources to understand RLHF (reinforcement learning with human feedback)
(1) "Asynchronous Methods for Deep Reinforcement Learning" (2016, arxiv.org/abs/1602.01783). Introduces policy gradients methods for deep learning-based RL as arxiv.org/abs/2203.02155…
Computational Linear Algebra for Coders.
100% FREE.
From the University of San Francisco's Masters of Science in Analytics program.
Essential for anyone looking to become a data scientist!
Link: github.com/fastai/numeric…
If your GPU supports mixed precision training, it's a good idea to enable it.
Fine-tuning a BERT model, mixed-precision can 4x training speed.
Shared the code here if you want to take a look: github.com/rasbt/deeplear…
Practical Deep Learning For Coders is one of the best FREE courses for those who want to learn Deep Learning.
• Build, train, and deploy models
• CV/NLP
• Tabular analysis
• Random forests and regression
• PyTorch, Fast AI, Hugging Face
course.fast.ai
GPT in 60 Lines of NumPy -- a simple yet complete technical introduction to the GPT as an educational tool.
jaykmody.com/blog/gpt-from-…
I love this! It reminds me why coding in Python can be so much fun!
🗨️ The most discussed research paper this week was "Toolformer: Language Models Can Teach Themselves to Use Tools"
It shows that LMs can use APIs to teach and improve themselves.
Paper: arxiv.org/abs/2302.04761
🗨️ The most discussed research paper this week was "Toolformer: Language Models Can Teach Themselves to Use Tools"
It shows that LMs can use APIs to teach and improve themselves.
Paper: arxiv.org/abs/2302.04761
We have little mechanistic understanding of how deep learning models overfit to their training data, despite it being a central problem. Here we extend our previous work on toy models to shed light on how models generalize beyond their training data.
transformer-circuits.pub/2023/toy-doubl…
📢Delighted to share Recurrent Interface Network (RIN), a new model/architecture that improves performance of state-of-the-art UNet-based image and video diffusion models (on pixels) while achieving 10X efficiency gain.
arxiv.org/abs/2212.11972
Joint work w @ajabri, @fleet_dj
2K Followers 858 FollowingAssistant Professor, University of Toronto & Vector Institute.
Probabilistic Machine Learning, AI Safety, AI Governance.
Prev: PhD & Rhodes Scholar at Oxford.
10K Followers 3K FollowingAddressing the world's most pressing challenges with science and technology. University of Cambridge, Department of Engineering.
2K Followers 879 FollowingNo longer here, emigrated to bluesky.
Professor of Statistical Machine Learning at @UniofAdelaide, @TheAIML. Previously at @OxCSML, @GatsbyUCL, @BristolUni.
18K Followers 4K FollowingAI professor.
Deep Learning, AI alignment, ethics, policy, & safety.
Formerly Cambridge, Mila, Oxford, DeepMind, ElementAI, UK AISI.
AI is a really big deal.
6K Followers 173 FollowingWe study mathematical principles of learning, perception & action in brains & machines. Funded by Gatsby Charitable Foundation.
Also on bluesky & mastodon.
3K Followers 4K FollowingTeach @OSA_BSS .Co-founded @BloodAid.
Tweets about business, research, cricket, football and everything in between.
People have allegedly died of my dad jokes!
1K Followers 1K FollowingByteDance Seed @ByteDance_Seed | Senior Research Scientist working on LLMs | prev. @oxcsml @UniofOxford, @amazon, @apple, @bloomberg
All opinions are my own
25K Followers 1K FollowingFind me @[email protected] Professor at @OxCSML, @oxfordstats and Research Director at @GoogleDeepMind. All opinions are my own.
4K Followers 2K FollowingDirector of Research in Machine Learning at the University of Cambridge, Head of Safe and Ethical AI at The Alan Turing Institute, MBE
6K Followers 2K FollowingSenior research scientist at @LosAlamosNatLab. Former prof at @ucl
and @UTAustin. CogSci, AI, Comp Neuro, AI for scientific discovery
Also @profdata on Bluesky
338 Followers 671 Followingmachine learning researcher in video and audio compression at qualcomm ai research amsterdam. electronic music is my second life. he/him
11K Followers 762 FollowingProfessor in Machine Learning, Gatsby Computational Neuroscience Unit
Research Scientist, Google DeepMind
https://t.co/JJNqmn7jak